Project Summary This K25 proposal is for a five-year mentored training program to enable Dr. Ariana Anderson from UCLA to transition from a statistician to an independent investigator, with an active research plan in identifying functional biomarkers of cognitive decline and aging. Dr. Anderson's proposal includes a comprehensive training program involving formal training in calibrated fMRI, fMRI data collection, neuropsychological assessments, volunteering, directed readings, coursework, mentoring, conferences, and career development, which are designed to complement and complete the applicant's previous training in mathematics and statistics. The proposed research addresses an important issue that affects nearly all fMRI studies; the modeled blood-flow response to neuronal stimuli is assumed to be constant across ages, genotypes and diseases, even though we know that this assumption is categorically false. This lowers the statistical power of fMRI studies, increases necessary sample sizes, and introduces bias into fMRI studies of disease and aging. Moreover, a poor understanding of hemodynamic change with age in healthy patients makes identifying biomarkers of unhealthy aging difficult. We will use cerebral blood flow measurements (ASL), hypercapnic and hemodynamic changes, along with genetic risk factors, as biomarkers to predict future cognitive decline. The relationship between the hemodynamic response, cognitive decline, aging and disease will be unraveled through the following three aims. Aim 1.) Create age-corrected hemodynamic response functions, after adjusting for cardiac and respiratory artifacts recorded during scan-time, so that future age studies can use age-corrected models of blood flow. We will estimate this in subjects with and without genetic risk for Alzheimer's disease. Age-corrected hemodynamic response functions will reduce bias and increase statistical power (increase the sensitivity, and/or reduce the required sample sizes). Aim 2.) Evaluate whether age- abnormal hemodynamics predict abnormal cognitive ability. Modeling normal hemodynamics will allow us to identify abnormal hemodynamics, creating biomarkers for specific diseases such as vascular dementia. Aim 3.) Create a new HRF model to account for age-related CBF changes, using calibrated fMRI. Abnormal hemodynamics may better predict which patients are more likely to experience cognitive decline, leading to earlier treatment.